I found a video on YouTube the other day covering a topic I have been contemplating. The topic? Joining Beams. Or more specifically, the topic of research is how to join beams at the ends to extend a beam. Or, maybe better put, what is the cheapest, easiest, most robust way to join beams?
Context of the research is for 2 areas of trailer engineering: 1) Attaching a gooseneck; 2) Adding a dovetail. I have seen and used various methods, so in this article we will look into some research on the technical side of how to join beams.
In trailer engineering at Syntheses we look for ways to do things better. Learning is awesome, so we start with what others do, then engage R&D (Ripoff & Duplicate) for analysis. There is often a better way to do things, so we look for good ideas and opportunities as a path to greater learning and continual improvement.
The Video That Caught My Attention
The video is titled "The Correct Way to Make Welded Splice Joints in Chassis Members." It is not exactly the conditions I am thinking about to join beams, but it is close. What can we learn?
While the video is a little laborious for someone unfamiliar with the math, you do not have to be an engineer to understand it. I think the author did a great job in touching the subjects, while not getting bogged down in engineering details. Geek out just a little on the math, then explain it. He delivers the concepts well.
Here is the video. I wish to give a big Thank You to the author for the work that went into it.
After watching the video a couple of times, I think he has done a commendable job with the engineering, and I extend a heartfelt THANK YOU for the effort.
Throughout the video, he poses questions about the optimal shape for splicing (joining) a beam. Overall I agree with his assessment from a purely theoretical standpoint - there should be no reason to make a diagonal, or to use steps.
I think his look at wood beams for comparison is insightful. I had not considered that years of wooden beams would influence thoughts for metal beams still today. Especially since wood beams have very different techniques for connecting - not welding like for steel beams. And, no one alive today was around before steel beams were common. However, the two materials and their respective approaches are not so dissimilar. (We will talk about that more below.)
Answering How To Join Beams
I felt the video was missing a clear answer to the original question. He did give a theoretical answer, but theory is not usually the full story. Practical circumstances will often change a theoretical answer.
For joining beams, we must also look at the reality of material and welds, because welding alters the material.
Analysis in the video assumes a homogeneous material, yet in reality, things change near a weld – in geometry and for material properties. First, welds do not penetrate perfectly through the material in a homogeneous manner. Then, there is always a dip near the edge of a weld, usually small and inconsistent, which also changes the geometry. Finally, the material temper changes near a weld, which changes the strength properties, and that affects how to join beams.
Though we like to think otherwise, the practical nature of the welding process is not homogeneous. Instead, the material ends up with gradients of strength and brittleness with slight bumps and valleys throughout the areas near a weld.
What does that mean? To get a true answer about joining beams, our theory would have to analyze through and including the post weld material gradients. Yes, both in material state gradients from heat disruption, as well as in material fill, and geometry. This is especially true when the welding rod material differs from the parent material. (Usually it does.)
Can we even know this information when we join beams? A thorough analysis is exceptionally difficult to do, so we normally handle it with proper processes and safety factors.
The video definitely takes Safety Factor into account in calculations. That is great, but it does not tell us how much of the safety factor we use when we make a weld in an otherwise homogeneous beam.
Service Conditions In Beam Joining Decisions
A big consideration as we join beams is the intended use. A static load beam that does not see load variation is pretty rare - especially with vehicles. So how much vibration do we expect? How much oscillation in the load? What is the amplitude?
The video mentions the beam in a vehicle, which will usually have complex loading over time. The same is true for a trailer frame. In these situations the potential effect of weld distortions are greater - because welds and fatigue don't play nice together.
True, this is highly material dependent. Aluminum is much more sensitive to fatigue than steel, for instance. However, no matter what material we choose, failures near welds are far more frequent than failures in areas that are free of weld interaction. I have worked on many situations analyzing these kinds of failures.
Without going into detail, the way a beam is used plays a major role in how we should join the beams.
Factory Processes / DIY Processes
Since the video does touch on it, we can point out that in a factory you have a lot more control in joining beams. Here are a few points to consider.
- First, In a factory they know the material, exactly, so they can choose a proper weld fill material. In DIY and aftermarket, the exact material of the original beams is often unknown. That means we usually join beams with whatever wire is in the MIG or whatever rod we have lying around.
- A factory usually has a lot more control for welding. Robotic welding is fantastic for consistency in power, speed, feed, and position. In DIY, the hand held - and usually less experienced hand - can't weld like a robot.
- In a factory, after welding they can heat treat the entire beam, like in an oven, to normalize the material and make the material properties closer to homogeneous. When we splice to extend a beam, it will never be truly homogeneous, but heat treatment after welding can drastically minimize the effects of a weld. We see this with bicycle frames where the parent material is quite thin. If bicycle frames are not heat treated after welding, they will fracture fairly easily. From a practical standpoint most people in DIY do not have access to that kind of equipment.
- In an environment of fatigue, crack propagation often starts, and propagates in the areas of previous heat distress. Most beam failures occur near welds, because the material properties have changed there. That allows a crack to propagate through and near the weld much more easily. Some say it is from the brittle areas internal where they join beams. Others say it is due to discontinuity in the material. I don't know the molecular reasons, but I definitely know the effect.
For DIY (and professionals not in a factory), I agree completely with the practicality of those who join beams - namely the independent trailer fabricators that must do the work, AND live with the results. Skill and care matter.
In college I worked as a mechanic. The boss told me "As a mechanic, you must be cleaner than a doctor." I pondered that for a minute, then because I looked puzzled, he finished. "Doctors get to bury their mistakes, you have to eat yours." That stuck with me, and it is certainly true of trailer builders that must live with / or eat the choices for how they join beams.
To combat these items above, especially crack propagation, having stresses spread farther in the beam connection can reduce the effect. Here are some examples.
Geometry To Extend A Beam
An angle cut through the beam gives more area of weld. Yes, that means more area affected, but it can also mean spreading it out. Unlike the conclusion in the video, I will argue that in practice, the geometry we choose does matter.
Another important aspect in failures is the direction of forces. For a beam in bending, the forces are along the beam, parallel, if you will, to the length of the beam. When extending or joining a beam, we generally want to avoid welds that are perpendicular to the direction of the forces. The discussion about halfway through the article from Mechanical Elements about welding things to trailer beams covers this concept in detail.
In the simple case of the video introduction image, the beam is cut at an angle (looking at the side), but straight across the top. (Cut 2 below.) Unfortunately, an angled side of the beam has less of an effect for spreading the stress load. If the angle cut is the opposite direction (Cut 3 below), or even better as a compound angle (Cut 4 below), the effect of spreading the weld is greater.
In the image, from left to right: 1. Straight cut; 2. Angle cut (side view); 3. Angle cut (top view); 4. Compound angle cut (angled both top and side views).
Comparative Reactions
Here are comparison results for the beams. For reference, we consider a beam with a moment load, bending in a vertical orientation. Obviously, these results need reorienting for loads in a different direction.
- A Straight Cut minimizes the cut perimeter, and therefore minimizes the weld length. At the point to join beams, all the weld is in the same plane, perpendicular to the forces of beam bending action. (Worst case.)
- Side View Angle Cut has a similar effect as the Straight Cut for the top and bottom surfaces, which are the most stressed for a beam in bending, so is only minimally stronger than the straight cut.
- Top View Angle Cut has more surface area for weld along the top and bottom surfaces, therefore it spreads the stress at the weld more. Yes, there is more heat affected area, so it is a bit of a toss up as to how much the strength will increase over the straight cut. However, to extend a beam, this is stronger. (Assuming good welds.)
- Compound Angle Cut (double diagonal if you prefer to think of it that way) is potentially stronger than the others simply because it spreads the weld around a longer path, and none of the path is perpendicular to the force directions.
The extent of the strength increase between these is arguable, because none of the welds above will approach the strength of a continuous beam.
Note, the above considers bending. Pure sheer is a simpler case. In general, more angle from the sheer direction puts less stress on the weld, so we will not cover that.
I believe the video gives us a good argument about why it does not matter how we join beams, but the missing factor is practical. From experience, I know angled welding is stronger than anything straight across a high stress surface. I also know the path to success when joining beams is additional material at those surfaces to minimize stress at the welds. We will see that later.
Theoretical Engineering / Practical Engineering
There are 2 really important engineering approaches. First, theoretical engineering (all the math and theory). Second, practical engineering. Both approaches are important, and they are not independent.
Theoretical engineering without the practical only works in school. Products do not make it in the market if the design is strictly theoretical.
Practical engineering is harder to describe, so we wrote a full article, about the practical side of engineering, and using experience to apply it. Please read that, because it also has examples in trailer engineering and bicycle design.
Looking around, it is pretty easy to see stuff that people say is "Over Engineered", and many trailer designers who join beams fit that paradigm. Truly, we can go overboard on the practical side by ignoring the theoretical, but that misses the optimization we hope to find as we look for the best ways to extend a beam. The best approach is to use both.
Think about it. Fractures are usually near a weld where the heat affected metal is not supported by the added weld mass. To know this and design for it is practical engineering. To know the beam to start with is theoretical engineering.
Practical Compensation for a Theoretical Problem
Part of the answer for the question of the video, is definitely in changes to material properties caused by welding. Basically, as we join beams by welding we are placing an interruption in the "flow" of stresses along the beam.
This is not unlike the wood beams of old. To join beams of wood, it is most practical to create some sort of overlapping joint - pin joint, finger joint, dovetail, etc. These increase the surface area to join, and at the same time create a path that is much more difficult for a crack to follow.
This is a geometry change. We can do the same with weld connections as we join beams.
The most common method with metals is to add stress plates over the joint. To extend a beam, we usually join beams with weld, then use plates over the weld seam. Prep the beam ends with chamfers to allow weld fill, then weld, then grind the weld areas flat before applying the stress plates.
Stress plates like in this image are common, and they do a great job of creating a complex joint that is hard for a crack to follow. They strengthen the areas of the weld in the beam, but they also introduce more stress with more welding.
One thing to help are corners turned diagonal. And, it helps even more to elongate the diamond shape. (To a point.) See the diamond shape side plates in the image.
One important thing to consider: Where the stress is greatest. This is the farthest faces from the neutral plane when in bending, which are the beams we are discussing. If that is where the stress is greatest, then a priority is to mitigate interruptions there.
Using Stress Plates To Spread The Load
Stress plates, like those in the image, help ONLY if we avoid welding on the top and bottom beam surfaces. So, in the image you will note the Blue representation of welds. (Colors only to illustrate the various areas.) They are only on the sides of the top and bottom long plates, not on the top or bottom surface of the beam.
Do not weld across the top or bottom face of the beam. Yes, we do weld across the top and bottom when joining the beam pieces, but that is under the stress plate. The finished beam should not have visible weld across the top or bottom surfaces.
We can do something similar on the sides, but this is less important if the beam is primarily in bending vertically. Like a trailer frame main beam.
We must be super careful about adding extra stress with more welding. We can actually weaken the beam instead of strengthening it. Strength in the final beam is as much about where we don't weld, as it is about where we do weld.
PLEASE NOTE: This is one simple example to join beams. Your situation, with your forces and your beam parameters will likely be different, so treat each beam in a manner to address the specific needs.
The Keys To Join Beams
Whether you need to extend a beam like for a trailer, or if you need to join beams for some other project, these principles apply. The greatest success will come as we thoughtfully consider the beam application, how the forces will apply, and the effects with our material choice. We can extend a beam successfully by applying both theoretical and practical engineering.
Now, to answer the question of the video, yes, as we join beams - shape matters! When joining or extending beams, the geometry, loading behavior, weld placement, and fatigue mitigation matter, because they are all a function of the shape.
The beginning of the video shows several examples of joining beams. I smile every time I see it. In these cases, things like bridges are exceptionally overbuilt (for very large safety margins) because the result of failure is the unthinkable. A vehicle chassis or trailer frame less so.
In the old days, to extend a wood beam they would often wrap the joint with leather or metal to strengthen it. With welding we can do something similar - though soaking leather is not needed.
To join beams with Bolts - like some of those in the beginning of the video - is a whole different animal. Maybe we will discuss those in a future article? Who knows.
In the meantime, if you are looking for more examples or more specifics about how to join beams for a dovetail or gooseneck trailer, see the companion article on our DIY site, Mechanical Elements.
Have A Wonderful Day!